Microsoft Word Viewer

advertisement
Modelling of propagation environments inside a
Scattered Field Chamber
Magnus Otterskog (Ph. D. Student)
Dept. of technology, Örebro University
Örebro, Sweden
magnus.otterskog@tech.oru.se
Abstract—Measurements on communication devices for evaluation
of diversity and/or MIMO concepts must be made in the real
environment or in a model of the real propagation environment. As a
first step towards a complex, physical environment model the plane
wave angular distribution, incident on the device under test (DUT), is
altered in order to model plane wave angular distributions of real
environments. Here a specialized Reverberation Chamber called the
Scattered Field Chamber (SFC) is used to create a source of Rayleigh
faded plane waves and a shielded, anechoic box with apertures is
used to alter the angular distribution of the plane waves incident on
the DUT. According to the measurements made, the model seems to
be able to produce data that show agreement with measurements
made on real propagation environments.
Keywords-propagation environment;
reverberation chamber; diversity;
radiated power [5] and the receiver sensibility [6] of cellular
phones for the last few years. These measurements can be
made both with and without the presence of an artificial human
head. However, in order to evaluate diversity and MIMO
concepts on communication devices one must test the device
in the real environment or in a model of the real environment.
Since measurements of real propagation environments [7][8]
show that the environments sometimes are far from isotropic
we need to create a specialized reverberation chamber that can
have different features than the traditional reverberation
chamber, for example a non-isotropic distribution of plane
waves that incident on the DUT. For this purpose we have
created the Scattered Field Chamber (SFC).
modelling; simulation;
I. INTRODUCTION
The Reverberation Chamber has been used in the EMC
(Electromagnetic Compatibility) community for several years
and it has been used as a military test equipment for decades.
The Reverberation Chamber [1] is a complex cavity made by a
shielded box, usually rectangular, containing some movable
conducting structure called stirrer, see Fig. 1. The cavity
should have a size that is in the order of 5-10 wavelengths in
any dimension to support the assumption that the cavity is
overmoded for the frequency that is used. Thus, every chamber
has a certain lowest usable frequency that is highly connected
to the chamber’s size and the uncertainty that can be accepted.
Another important factor is the stirrer and how it is constructed
[2][3]. The stirrer should, as it moves, change the boundary
conditions for the fields inside the cavity and thus create
different mode structures inside the chamber. An efficient
stirrer will create a larger number of different mode structures
than an inefficient stirrer and this will give a larger number of
independent measurement samples and a smaller statistical
uncertainty. The amplitude of one component of the electric
field inside the chamber measured for a large number of
different stirrer positions follows a Rayleigh distribution. The
received power in an antenna will therefore follow the
exponential distribution [4] and this is similar to a real
propagation environment without presence of a direct
propagation between the antennas. The environment inside the
ideal reverberation chamber is also statistically isotropic and
statistically uniform. Different versions of the traditional
reverberation chamber have been used for measurements of the
II. MEASUREMENTS ON REAL ENVIRONMENTS
Measurements on real propagation environments [7][8]
show that the distribution of plane waves that incident on the
DUT can vary very quickly. For example in an indoor office
environment the received power can vary more than 20 dB for
incident angles that is changed only a few degrees. Outdoor
measurements in urban environments show variation of the
same size but not as rapid angle dependence. Measurements
like these give us the guidelines for the environment that we
would like to physically model.
III. SHIELDED ANECHOIC BOX MEASUREMENTS
A. Small chamber
A combination of the Reverberation Chamber and the
Anechoic Chamber is used for the model. A small shielded
anechoic box is placed inside a larger Reverberation Chamber,
see Fig. 1. The inner walls of the box are covered with
absorbing material and the angular distribution of plane waves
that incident upon the DUT, placed inside the box, is
controlled with apertures in the walls of the box. Here the
reverberation chamber is used as a source of Rayleigh
distributed plane waves [9] and the shielded box with its
apertures is the model of the plane wave environment. We
show the relative received power as a function of rotational
angle of the receiving antenna in Fig. 2. The measurements
shown in Fig. 2 were conducted at 2.45 GHz in a small
reverberation chamber of dimensions 1.0×0.5×0.5 meters. The
shielded box, that covered the receiving antenna, was a
Figure 1. A small SFC with a shielded anechoic box with apertures.
Figure 2. The relative received power as a function of rotational angle of the
receiving antenna placed inside the shielded box.
Figure 3. The large shielded anechoic box with broadband horn antenna.
distance to the aperture. The aperture was a circular
aperture of 60 mm diameter and this is shown in Fig. 3. The
broadband horn antenna was rotated with 22.5 degrees steps
between 0 and 180 degrees giving 9 different directions. The
standard gain horn with 22.5 degrees steps between 0 and 90
degrees. The aperture is placed in the 0 degree direction and at
a distance of 53 cm from the aperture of the antennas. The
relative received power was measured for 51 frequencies
between 1.7 and 2.6 GHz. The result of a measurement at a
single frequency for the first horn antenna, the broadband horn
antenna with an approximate 3 dB beamwidth of 60 degrees
for the used frequency, is presented in Fig. 4. The same thing
for the other antenna, the standard gain horn antenna with an
approximate 3 dB beamwidth of 30 degrees, is presented in
Fig. 5. As we can see the reduction rate of plane waves seems
to be faster for the standard gain antenna and thus the effect of
the directivity of the receiving antenna is significant. Since
antennas of finite directivity are used in the measurements the
true reduction rates of plane waves from different directions
are underestimated.
cardboard box covered with aluminium foil and quadratic
apertures of size 8 cm in the directions 0° and 180°. The
interior of the box was covered with absorbers of type
ECCOSORB LS16-22. The receiving antenna was rotated in
45 degrees steps between 0°and 180°. The receiveing antenna
was a WLAN antenna with an approximate directivity of 8
dBi. This directivity will give a large 3 dB beamwidth and the
antenna will receive a lot of energy directly from the aperture
even though it is pointed away from the aperture. This will
affect the measurements and the directional properties of the
aperture will be underestimated.
B. Large chamber
In order to examine the effect of the finite directivity of the
used receiving antenna we have also performed measurements
as the ones above but this time in a large reverberation
chamber (37 m3) with a large shielded anechoic box (1 m3)
inside. This time two types of horn antennas with different
directivity were placed inside the anechoic box with a constant
Figure 4. The relative received power as a function of rotational angle of the
receiving broadband horn antenna.
Figure 5. The relative received power as a function of rotational angle of the
receiving standard gain horn antenna (only 0 to 90 degrees).
IV. FIELD STATISTICS
The statistical distribution of the plane wave amplitudes
that is received by the antenna is shown in Fig. 6. The
cumulative distribution function (CDF) of the measured data
are plotted together with a best fit CDF of a Rayleigh
distributed variable. The measurement data are taken for the
broadband horn antenna when it is pointed 22.5 degrees away
from the circular aperture and the frequency is 2.6 GHz. Fig. 7
shows the same distributions but this time with logarithmic
scales in order to enlargen the area for small signal levels.
Figure 7. The CDF of measurements compared with best fit theoretical
Rayleigh distributed variable, logarithmic scale.
important factors of how large variation and how fast
variation with angular shift that you can obtain in the
measurements. The use of absorbers with finite attenuation and
receiving antennas with finite directivity will underestimate
both the possible maximum reduction of plane waves and also
the maximum rate of reduction with angular shift of the
receiving antenna.
VI. FUTURE WORK
Future work includes numerical simulation of the plane
wave distribution inside a shielded, anechoic box with
apertures as described above. Modelling of the statistics of the
incoming plane waves, Rayleigh, Rice or Gauss distributed,
has already been shown [10] but will need some further
investigations. Methods to control the polarization of the
incoming waves and the time spread must also be invented in
order to model different real environments. In the end, real
field tests on communication devices in different environments
will be compared with tests in the SFC and the correlation will
be examined.
ACKNOWLEDGMENT
I would like to thank Olof Lundén at the Swedish Defence
Research Agency for valuable support when performing the
measurements.
REFERENCES
Figure 6. The CDF of measurements compared with best fit theoretical
Rayleigh distributed variable, linear scale.
V. CONCLUSIONS
The distribution of plane waves that incident upon the
DUT can be controlled inside a SFC. The amount of reduction
is high when a shielded, anechoic box with apertures is used to
control the distribution. Reduction rates of about 20 dB can be
obtained. The absorber quality and antenna directivity is
[1] Bäckström, Lundén, Kildal, “Reverberation Chambers for EMC
Susceptibility and Emission Analyses”, URSI the Review of Radio
Science 1999-2002, 2002.
[2] Wu, Chang, “The effect of an Electrically Large Stirrer in a ModeStirred Chamber” IEEE Transactions on electromagnetic compatibility,
vol. 31, No. 2, May 1989.
[3] Lundén, Bäckström, “Stirrer Efficiency in FOA Reverberation
Chambers. Evaluation of Correlation Coefficients and Chi-Squared
Tests” Proceedings of the 2000 IEEE International Symposium on
Electromagnetic Compatibility, Washington DC USA, 2000.
[4] Hill, “Electromagnetic Theory of Reverberation Chambers”, NIST
Technical Note 1506, National Institute of Standards and Technology,
December 1998.
[5] Kildal, Carlsson, “TCP of 20 mobile phones measured in a reverberation
chamber”, Bluetest report 2 (available from address: Bluetest AB,
Chalmers teknikpark, 412 88 Gothenburg, Sweden or from e-mail:
simon@kildal.se), Nov 2001.
[6] Orlenius, Serafimov, Kildal, “Procedure for measuring radiation
efficiency in downlink band for active mobile phones in a reverberation
chamber”, IEEE Antennas and Propagation Society International
Symposium 2003, Volume: 4, June 22-27, 2003, Pages 731-734.
[7] Kalliola, “Experimental analysis of multidimensional radio channels”,
Thesis for the degree of Doctor of science in Technology, Helsinki
University of Technology, 2002.
[8] Medbo, Berg, “Spatio-Temporal Channel Characteristics at 5 GHz in a
Typical Office Environment”, IEEE VTC Fall 2001, 2001.
[9] ] Svensén, “Reverberation Chamber as a radiation source”, The 2003
Reverberation Chamber, Anechoic Chamber and OATS Users Meeting,
Austin TX USA, 2003.
[10] Madsén, Hallbjörner, “Reverberation Chambers for Mobile Phone
Antenna Tests”, The 2001 Reverberation Chamber, Anechoic Chamber
and OATS Users Meeting, Bellevue WA USA, 2001.
Download